Skip to content

babynata/hydration-reminder-agentic

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 

Repository files navigation

Hydration Reminder Agentic

A fully personalized, agent-agnostic water-drinking reminder skill for AI assistants. Works on Claude Code, Claude.ai, Cursor, GitHub Copilot Chat, or any AI agent platform that supports skill instructions.

Not a one-size-fits-all plan. It interviews you first, then designs a custom daily hydration schedule with interactive feedback loops, progress tracking, and adaptive reminders.


Features

  • Personalized Consultation — Asks about your drinking habits, goals, health conditions, cup size, work schedule, and reminder preferences before creating anything.
  • Smart Plan Design — Adapts to office workers, gym-goers, creatine users, pregnant women, people who hate plain water, small-cup users, forgetful drinkers, etc.
  • Visual Progress Tracking — Every message shows today's total, a progress bar, and remaining amount.
  • Daily Interactive Loop — Remind → ask for intake feedback → dynamically adjust remaining reminders → bedtime summary.
  • Cross-Platform — Works on Claude Code (/loop, CronCreate), Claude.ai, or any agent platform. Honestly discloses platform limitations and provides fallback options (phone alarms, calendar, WaterMinder, Hydro Coach).
  • Bilingual — Full Chinese (喝水提醒) and English (water reminder) support.

Quick Start

For Claude Code Users

  1. Clone or download this repo:

    git clone https://github.com/babynata/hydration-reminder-agentic.git
  2. Copy the skill to your Claude Code skills directory:

    mkdir -p ~/.claude/skills/hydration-reminder-agentic
    cp hydration-reminder-agentic/SKILL.md ~/.claude/skills/hydration-reminder-agentic/
  3. Restart Claude Code or type /skills to reload.

  4. Say anything like:

    • "帮我设置喝水提醒"
    • "water reminder"
    • "我健身,吃肌酸,目标2升"

For Claude.ai / Other Platforms

Since these platforms don't support external skills directly, copy the contents of SKILL.md into a new Project, or paste it as a system prompt. The skill instructions will guide the AI through the full workflow.


How It Works

Phase 1: Consultation

The assistant interviews you with 4 key questions (background, goal, conditions, frequency) before designing anything. Even if you provide context upfront, it asks 1-2 follow-ups to fill gaps.

Phase 2: Plan Design

Based on your answers, it designs a schedule like:

你的定制喝水计划预览:
━━━━━━━━━━━━━━━━━━━━━━
🎯 日目标:2000ml(约8杯)
⏰ 提醒时间:09:00 / 12:00 / 13:00 / 17:00(肌酸) / 18:00 / 21:00 / 22:30
📱 推送渠道:当前对话窗口
📝 特殊备注:健身+肌酸,17:00肌酸+300ml水
━━━━━━━━━━━━━━━━━━━━━━

Phase 3: Deploy

On Claude Code, it can use CronCreate or /loop to schedule real recurring reminders. On other platforms, it provides a copy-pasteable schedule for your phone/calendar.

Phase 4: Daily Loop

Every reminder includes:

🥤 喝水提醒
━━━━━━━━━━━━━━━━━━━━━━
██████░░░░ 900ml / 1500ml (60%)

⏰ 本次:喝 300ml
🎯 还剩:600ml
━━━━━━━━━━━━━━━━━━━━━━

喝完后回复我数字,我帮你记录。

Phase 5: Summary

At bedtime (22:30), you get a full daily summary with achievement rate, segment records, and tomorrow's suggestion.


Supported Scenarios

Scenario How It's Handled
Office worker Reminders aligned with work breaks
Gym + creatine Dedicated creatine+water reminder, asks for training time
Pregnancy / doctor's limit Gentle tone, smaller amounts, medical disclaimer
Hates plain water Tea, lemon water, infusions all count
Small cup (200ml) More frequent reminders, exactly 1 cup each
Always forgets Aggressive morning push (3 pre-noon reminders)
Modify plan mid-day Adjust target, redistribute amounts, update preview
Skip today Respects request, shows current progress, resumes tomorrow
Stop reminders Confirms immediately, shows resume path

Trigger Phrases

Say any of these to activate:

  • 喝水提醒, 设置喝水, 帮我提醒喝水, 定制喝水计划
  • water reminder, hydration reminder, set up water reminder
  • 修改喝水, 更新喝水, 调整提醒
  • 停止喝水提醒, 关掉喝水, 暂停喝水
  • 查看喝水进度, 今天喝了多少, 喝水总结

Tech Spec

  • Skill format: Standard Claude Code SKILL.md with YAML frontmatter
  • State storage: Platform-dependent (file, memory, or database). JSON schema provided in skill.
  • Scheduling: Platform abstraction — uses whatever the host supports (cron, timers, loops, or manual fallback).
  • Language: Bilingual instructions; AI responds in the user's language.

Contributing

This skill was built using the skill-creator workflow with iterative eval-driven improvement. If you have ideas for new scenarios, better phrasing, or platform-specific adaptations, feel free to open an issue or PR.


License

MIT

About

Personalized water-drinking reminder skill for Claude Code / Claude.ai / any AI agent. Interview-based plan design, visual progress tracking, daily feedback loops. Bilingual.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors